INTRODUCTION: The association between physical activity and mood is of major importance to increase physical activity as a prevention strategy for noncommunicable diseases and to improve mental health. Unfortunately, existing studies examining how physical activity and mood wax and wane within persons over time in everyday life do show ambiguous findings. Taking a closer look at these studies reveals that the aggregation levels differ tremendously. Whereas mood is conceptualized as a three-dimensional construct, physical activity is treated as a global construct not taking into account its distinct components like exercise (such as jogging) and non-exercise activity (NEA; such as climbing stairs). METHODS: To overcome these limitations, we conducted an ambulatory assessment study on the everyday life of 106 adults over 7 days continuously measuring NEA via accelerometers and repeatedly querying for mood in real time via GPS-triggered e-diaries. We used multilevel modeling to derive differential within-subject effects of exercise vs. NEA on mood and to conduct analyses on the temporal course of effects. RESULTS: Analyses revealed that exercise increased valence (beta=0.023; p